The influence of nursing home characteristics on the accuracy of the long term care minimum data set

Abstract

In June 1998 nursing home providers began transmitting MDS data to the Health Care Financing Administration's national repository. By the end of 2000 the national MDS database contained approximately 36 million records. MDS data are used for casemix payment, to focus the survey process and to provide quality of care feedback information to nursing homes. In the near future, MDS data will also be used as part of the medical review (MR) system to ensure that Medicare beneficiaries are furnished appropriate services in skilled nursing facilities. The potential magnitude and implications of the decisions for beneficiaries that are and will be based upon MDS data underscores the need for the data in the system to be accurate. While this requirement may seem obvious there are, unfortunately, many indications that the data may indeed not be accurate (Harrington et al., 1996). The research suggests there are characteristics of the nursing home's organizational environment that may make it difficult to obtain accurate data (Mumford, Whetzel, & Reiter-Palmon, 1997; Phillips, 1995; Harrington, 1996; Abt, 2001). The purpose of this study was to determine the influence of nursing home characteristics on the accuracy of MDS data. This study was conducted in two phases. During phase one a model of nursing home characteristics that influence the accuracy of MDS data was developed. Ten MDS experts were interviewed and asked to identify nursing home characteristics that influence the accuracy of MDS. Nine characteristics of nursing homes were identified. The nine characteristics were used to construct a model for this study, based on Donabedian's (1980, 1982) components of structure, process and outcome. The structure component included the following facility characteristics: the number of beds & units, occupancy rate, ownership, staffing levels, used of per diem staff, a designated MDS coordinator, MDS coordinator characteristics (years of experience, training, employment status), and resident characteristics (cognitive impairment and medicare status). The process component of the model included the tenure of the MDS coordinator, the process used to complete the MDS and the turnover of the DON. The outcome component (dependent variable) was the accuracy (quality) of the MDS data.;The sample for this study was limited to fifty-one nursing homes in the state of Maine. Three nursing home characteristics were found to have an association with the accuracy of MDS data. MDS data were more accurate in facilities that had a designated MDS coordinator. A positive relationship was found between the level of experience of the MDS coordinator and MDS data accuracy. Data appeared to be more accurate in facilities where the MDS coordinator had more experience. The accuracy of the MDS data was lower in facilities that use per diem staff. The findings from this study suggest more research that examines the impact of the MDS coordinator and the necessary qualifications for this role needs to be conducted. Further, the alternative methods to assess MDS data accuracy are needed. Lastly, consideration must be given to refining and decreasing the complexity of the MDS.